Post sponsored by NewzEngine.com

Source: MIL-OSI Submissions
Source: University of Auckland

Robots might be good at welding, packaging and assembling cars, but creating robots that have the sensitivity of a human hand, capable of cradling an egg or picking a berry without swashing it, remains a challenge.

This is a challenge that Dr Samuel Rosset and his team at the Auckland Bioengineering Institute (ABI) aim to address by developing a soft sensor and sensing algorithms, that could make robots more sensitive to touch – to better equip them to, or example, pick fragile fruit without bruising them. They have received $200,000 in SfTI (Science for Technological Innovation) to do help them do so.
There are many things about humans that we, as humans, take for granted. “We can shake someone’s hand without inflicting pain, because of the soft and compliant nature of our skin,” says Dr Rosset.
“Our hands have the ability to conform to fit within or around the shape of object, such as someone else’s hand; we instinctually know what pressure to apply, which isn’t too soft, but not too hard. But while robots are good at manipulating rigid materials, interaction with soft objects is much more difficult – something that is simple for humans, such as shaking hands, is a real challenge for robots.”
Robots are increasingly seen as a solution to the labour shortage in the New Zealand fruit-picking industry, but despite progress in the development of key robotic components such as vision systems that can identify when a fruit is ripe, creating robots that can pick the fruit without bruising it as a human would is an ongoing challenge.
Dr Rosset and team aim is to develop a skin-like material that provides a soft compressible interface with objects, and that can sense how much force is being applied – so it can be programmed, for instance, to cradle an egg without crushing it.
Although many sensing skin concepts have been developed, they rely on micro structured layers, or organic transistors that are complex and expensive to produce. Moreover, they are aren’t compressible. To respond to the shape of an object as our hand does when it picks a berry, requires a sensing technology that is easily compressible, so it can conform to the target and distribute the pressure.
Moreover, they aim to equip the technology to detect location – so it can sense where an object is, and how much force should be applied where. “So that it’s a bit like our skin, which gives information on pressure and location: if someone/something pokes your skin your nerves sends your brain information on where and how strongly they’ve poked you.”
Dr Rosset and his team will draw on the extensive knowledge developed by the Biomimetics lab at the ABI, which is world-renowned for its stretchable sensing technology, but adapting the technology that measures ‘stretch’ to measure compression – how much pressure is being applied to an object.
This level of sensing technology would typically require a complex array of electrodes with a lot of wiring. What is particularly unique and innovative about Dr Rosset and teams’ research is how simple they aim to keep things.
Instead of relying on a complexity of electrodes they aim to use only one electrode, and a smart sensing algorithm, that allows for new ways of ’electrical probing’.  This ultimately will mean the technology will be easy and cheap to manufacture
“So we want to keep the sensor simple, but by identifying more complex signals, we can enable the technology to detect complex information. It would allow us to give robots a skin, and a sensitive skin, one that can locate the berries and other fruit and pick them without bruising them.”

MIL OSI